AI Sales Automation Pricing: What Saudi SMEs Should Expect

Saudi SMEs should budget for AI sales automation as a complete workflow: software or usage, sending infrastructure, approved data preparation, human review, integration and reply handling. The cheapest advertised unit can be a small part of the actual operating cost. Compare proposals using the same audience, activity definition and responsibilities.

This guide provides a pricing model and purchasing questions rather than an invented Saudi market average. It also records the public Post commercial model reviewed on 26 October 2026. Use the B2B outbound service guide to distinguish a product discussion from a scoped prospecting audit or implementation service.

What Post publicly states about pricing

On 26 October 2026, the live Post homepage advertised 100 credits for new users, one credit per email, pay-as-you-go top-ups and no monthly or annual subscription. The reviewed public page did not establish a price per purchased credit. Check the current account terms and purchase screen before budgeting or buying.

Those statements describe the published model, not every billing rule. Confirm what counts as an email for charging purposes, when credits are deducted, and how failed attempts, retries, regenerated drafts and unused balances are treated. Do not infer these details from the phrase one credit per email.

Keep product usage separate from services. A prospecting audit, campaign preparation, Arabic review or custom CRM work may involve additional scope and cost if agreed. A product credit balance does not establish that those services are included, and this guide does not present a quotation for them.

Compare pricing units before comparing totals

Common pricing units and the questions they create

Unit	What to clarify	Comparison risk
User or seat	Named users, administrators and approval roles	A low base price may exclude people needed to operate
Email or credit	Billable event, retries, limits and balance treatment	A generated draft may not equal a sent or accepted message
Connected account	Mailbox count, permissions and supported providers	Extra connections may add cost without useful capacity
Data record	Source, permitted use, freshness and acceptance	More records do not necessarily mean eligible prospects
Managed service	Research, review, execution and response responsibilities	A bundled fee may hide exclusions or duplicated work
Implementation	Integrations, testing, handover and change requests	A low setup quote may assume clean systems and data

Ask each supplier to price the same defined workload. Do not compare a self-service tool fee with a managed campaign fee as though they buy the same result. Identify which tasks your team must supply in each case and cost them consistently.

Use the contracted unit in the forecast, then track actual consumption. If a proposal uses a proprietary credit system, request a worked example with your intended operations. A simple conversion table is more useful than an unexplained monthly allowance.

Model message volume from sequence behaviour

Start with eligible people and the maximum reviewed sequence, then account for the rules that stop later messages. In an illustrative campaign, 500 people could each receive an initial message and up to two follow-ups, giving a theoretical maximum of 1,500 messages. That maximum assumes all remain eligible for every step.

Suppose ninety leave the sequence after the first message because a reply or stop condition ends it, and another fifty leave after the second. The planned flow would then be 500 first messages, 410 second messages and 360 third messages: 1,270 in total. These are fictional planning assumptions, not an expected response pattern.

The count is not automatically the credit bill. The supplier's definition of a billable email still controls charges. Bounces, retries and other events may have specific treatment that must be confirmed. Keep the message forecast and the charging rules as separate parts of the model.

Budget for the human work around the product

List audience review, evidence checking, Arabic and English editing, approvals, reply handling, qualification and reporting. Record the time required for a representative batch rather than assuming that generation eliminates these tasks. A short pilot can reveal which parts dominate effort.

Include exception work. Incorrect identities, unsupported claims, ambiguous replies and failed integrations require attention. If the budget assumes every record succeeds on the first pass, it will understate cost precisely where the workflow becomes difficult.

Separate cash expenditure from internal capacity. Existing staff time may not create a new invoice, but it can displace other work. The manual versus assisted comparison explains how to cost accepted outputs fairly. Do not count released hours as payroll savings unless an actual expense changes.

Work through an illustrative launch-month budget

Consider a hypothetical SME planning a controlled launch for 500 eligible people. Assume SAR 3,000 for one-time setup, SAR 1,500 as a planning allowance for software and sending, thirty hours of research and review at SAR 120 per hour, fifteen hours of response handling at the same rate, and SAR 600 for integration maintenance. These are arbitrary modelling inputs, not Post prices or Saudi market benchmarks.

The budget is SAR 3,000 + SAR 1,500 + SAR 3,600 + SAR 1,800 + SAR 600 = SAR 10,500. Dividing by 500 gives SAR 21 per eligible person prepared and handled under this scenario. It is not cost per lead qualified, meeting held or customer won.

Show the one-time setup separately so a later recurring-month comparison does not assume it repeats unchanged. Also show the actual cash payment schedule. If an annual commitment is proposed by another supplier, do not compare its annual cash outflow with one month's usage from a different option without explaining the horizon.

Replace every assumption with a quote, invoice or measured workload as evidence becomes available. The purpose of the example is to reveal omitted categories and sensitive drivers. It is not to suggest that an SME should spend this amount or expect a particular commercial outcome.

Test the assumptions that can change the decision

Use the launch example to identify the largest uncertainty. If research and review take fifty hours rather than thirty at the same assumed SAR 120 rate, the additional twenty hours add SAR 2,400. The total becomes SAR 12,900, or SAR 25.80 per eligible person. That is about 22.9% above the original SAR 10,500 budget, even though the audience and software allowance have not changed.

This sensitivity test gives the pilot a concrete measurement task: record review time and reasons for rework. It does not justify cutting essential review to preserve the original estimate. A narrower offer, better source material or clearer acceptance rules may address the cause more effectively than buying a larger credit balance.

If the one-time setup genuinely does not recur and all other illustrative costs remain the same, the next-period operating scenario would be SAR 7,500. Keep that conditional assumption visible. A new integration, additional language or changed process could create fresh setup work. Do not advertise the lower figure as an ongoing market price.

Reconcile usage, invoices and operating effort separately

Create three linked records: a usage ledger for supplier-defined billable events, an invoice ledger for actual cash charges, and a workload ledger for internal time and accepted outputs. A purchased credit balance is a cash commitment; consuming credits is usage. Recording both as a full new expense in the same cost model can double count the same purchase.

Ask the finance owner how to recognise prepayments and unused balances in the chosen management report. Keep the cash view visible even when an operating comparison allocates usage over several periods. The supplier's current terms determine whether a balance expires or is refundable; do not assign it a recoverable value without evidence.

Investigate unexplained differences before scaling. A billing discrepancy may involve a charging definition, a duplicate event, a delayed usage update or a reporting error. Preserve the event reference and request a reconciliation. This creates a specific commercial question for the supplier instead of an unsupported allegation about overcharging.

Compare self-service, assisted and managed arrangements

In self-service, your team may own audience preparation, review, sending setup and reply handling. An assisted arrangement may add implementation or campaign support while leaving commercial decisions with you. A managed service may take on more operation, but its scope still needs explicit boundaries.

For each proposal, create a responsibility table. Mark who supplies data, approves claims, verifies eligibility, edits Arabic, manages technical settings, handles responses and reports results. A task left blank is a future surprise, not a saving.

Check whether two suppliers are charging for the same work. A CRM partner and campaign agency may both include data mapping or reporting. Conversely, each may assume the other owns suppression or failure recovery. Resolve overlaps and gaps before comparing headline fees.

Ask which costs rise with scale

Some costs rise with messages, others with people, accounts, campaigns or complexity. More generated messages may be cheap while review and response handling remain constrained. A second language or a new industry can add work even if the contact count stays constant.

Test the cost curve at a modest expansion rather than multiplying the pilot result mechanically. Ask whether a higher tier, additional access, support package or integration change is required. Keep any quoted threshold tied to the supplier's current terms.

Model the cost of quality. If the acceptance rate falls as the audience broadens, more records and review hours may be needed per useful result. Include the rejected work and the reason. Scaling an unsuitable audience can increase spending without increasing commercial value.

Read commercial terms that affect flexibility

Check minimum purchase, expiry, refund, cancellation, renewal and unused-balance treatment where applicable. Confirm whether quoted amounts include taxes, fees and any currency conversion. Use the actual invoice and applicable treatment rather than assuming that every supplier presents prices on the same basis.

Ask what happens if the integration or campaign does not launch as planned. Can usage be deferred? Are implementation milestones accepted separately? What support is included, and what constitutes a chargeable change? These details can matter more than a small difference in the advertised unit price.

Keep the terms attached to the relevant date and version. A pricing statement from an old article or a previous account screen may no longer apply. A budget owner should be able to trace each material number to a current quote or a clearly labelled assumption.

Use meaningful unit economics after launch

Track cost per accepted record, relevant human conversation, meeting held and accepted opportunity where the data supports each measure. State the denominator and period. Do not switch from cost per person to cost per message simply because the latter looks smaller.

Connect those operating measures to the contribution-based ROI model as sales mature. Pipeline is a forecastable commercial asset, not collected profit. Use the pipeline contribution method to avoid assigning every existing opportunity to the latest campaign.

A low cost per message can coexist with a high cost per useful conversation. The right optimisation depends on the business decision. If the offer is wrong, buying more low-cost sending capacity will not fix the economics.

Negotiate scope and evidence rather than only price

Ask for a defined pilot, clear acceptance criteria and a transparent handover. A supplier may be able to reduce cost by narrowing the initial integration or audience without weakening an essential control. Do not negotiate away the work needed to make the workflow operable.

Separate optional enhancements from the minimum viable arrangement. Identify what can wait until the pilot proves value. Keep a change-control method so later requests receive a clear impact on price, timing and responsibility.

Use the platform buying framework to assess the complete decision. The cheapest quote is useful only if its assumptions match your actual needs and the business can fulfil the responsibilities left outside it.

Prepare a pricing request that produces comparable answers

Provide the eligible audience size, planned sequence, languages, approved offer, sending arrangement, CRM needs, review responsibilities and response capacity. Ask suppliers to identify their billable units and all exclusions. Request a worked consumption example and a clear distinction between current product capability and custom delivery.

For Post, request a demo or commercial discussion and confirm the current top-up terms in the live application. Build the budget from verified terms and your measured work. That produces a decision the SME can operate and review without relying on an unsupported market-wide price promise.